The destruction of C is guaranteed

> java is slowly dying
Not sure about that, and anyhow the JVM languages (so, Java + Scala + Kotlin + Clojure + ...) seem to be growing if anything.

Especially in the markets where the likes of Go and Rust users claim they should be the best option. Concurrent, distributed, big ... systems. Lots of data, lots of processing, lots of users, and a need for it all to work reliably.

GCd languages are excellent for concurrent distributed big systems, which is why Rust probably wont be picked up for that sector; go is GCd (with a faster GC algorithm than java AFAIK) so it could do well but lol no generics

Java isn’t good at that work natively. It’s the supporting infrastructure which has built up around it that makes it useable.

This often isn't really infrastructure software-wise. Much of it is newer stacks based on Scala Spark / Finagle / ... and software built with SBT and so on, or newer very much repeatedly updated Java stacks.

Even if your observation was correct -which I doubt, we don't live in a world where the advantages of the JVM count for nothing-, at the very least its "native" ability to build a good supporting infrastructure quickly would be remarkable.

OTOH this wasn't like "someone coded something and the JVM was just the first" type of situation. The JVM won like 20-30 years after a lot of people started to want more parallel computations, everyone also had a good shot.
OpenMP and OpenHMPP and CUDA and so on had a lot more support for a good while and some people believed and invested into these believing they would be used for performant local or distributed parallelism. But the JVM still won.

What does Rust have to offer over C?

obviously, YOU DON'T HAVE TO CARE ABOUT ALLOCATING MEMORY SO THAT'S GOOD, ISN'T IT?

>he doesn't know oracle killed java in February
F

You mean them moving everyone to OpenJDK or derivatives?

That only strengthened the JVM, I think.

Go has nice syntax and fast compile times.
But it has shit performance and a large runtime.

If you want a fast language you either go unsafe or do everything statically like rust.

I still like Go though, it's comfy in a way unlike other languages, but I wouldn't build half my userspace with it.

I'm fine with the JVM. It's java that's a pain to write.

Actually you do have to care about, ever more so than C because in Rust your program won't even compile if you don't.

But seriously. Who thought "Rust" would be a good language name?

Every time I see it I think it's going to be an id Software game.

i think rust is going in the right direction, but i hope the project dies because too many fucking LGBT COC people are behind that shit. Like just post your patches and don't bother people with your PC bullshit please

Rust either:
>Uses unsafe() everywhere, hence is just a worse version of C (just like C++)
or
>Doesn't use unsafe, and is slow as shit unless you write unreadable code to make the compiler find "happy paths", violating the principle of "don't do premature optimization"
medium.com/@shnatsel/auditing-popular-rust-crates-how-a-one-line-unsafe-has-nearly-ruined-everything-fab2d837ebb1

>pay massive for support
>get fucked whenever possible
>have to pay even more
in our team we started to transition because of their shenanigans
PL -> Postgres
Java -> c++ / c#

idk what others in the industry do.
it's annoying to review old code tho
but also a much needed cleanup

JVM's got its issues too. Not terrible or anything, but way too much baggage. If you're looking for the biggest platform support it's your best bet, but if you're just targeting the main ones, .net core's actually better.

GC'd langs are absolutely terrible for concurrent distributed big systems unless you have separate heaps like Erlang.

At least with C, I know it's limitations, and often the only limitation is the hardware.

Everything I try to write code in C++/Rust, I either end up writing C89 style code (using "unsafe"/no STL/etc.), or I have to write a massively over-complicated macro/template to solve my basic problem.

I'd rather program in C11 with some Preprocessor/Makefile magic. It's much easier to reason about my code that way anyway (which is supposed to be the selling point of Rust/C++).

You can't even create bitfields in Rust, and C++ forces you to bloat your code with constructors/destructors/templates everywhere when you use STL.

I personally also wouldn't want much to do with Oracle itself - heard many bad things and little good about them.

OTOH what's the issue with just using one of the OpenJDK fork distributions? Ultimately, Amazon, SAP, the AdoptOpenJDK consortium and many more offer it. Seems easier to me than trying to move to C++ and C# (the latter also wasn't shenanigans-free, right?).

Rust is a great language.
No it will not replace c in the next 30 years.
Considering the ammont of Fortran still used, C will not be replaced in the next 60 years.

I think you can objectively argue that C has replaced Fortran, though... (C even has an optional "fortran" keyword)

>Doesn't use unsafe, and is slow as shit
unsafe{} doesn't magically make code run faster, and the hand-optimizations it allows are usually minor.

>violating the principle of "don't do premature optimization"
If you're comparing profiling results of changes to already-written code, that's not "premature optimization", its just optimization.

>medium.com/@shnatsel/auditing-popular-rust-crates-how-a-one-line-unsafe-has-nearly-ruined-everything-fab2d837ebb1
>In case fuzzing turns up no bugs, eyeball the unsafes and try to check them for memory errors.
That sounds like the unsafe{} keyword is doing it's job perfectly. The whole point of it is to limit the locations where memory bugs could occur down to a small handful of searchable blocks.